exa-search

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow uses the Exa MCP server to perform web/code searches (public web pages and other third-party text), which the MCP tool then returns as readable prose into the agent’s LLM context—an outsider-authored free-text source via livecrawl/search results.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires running "npx -y exa-mcp-server" to start the Exa MCP server (which fetches and executes the exa-mcp-server package from the npm registry, e.g. https://www.npmjs.com/package/exa-mcp-server) at runtime to provide the web_search_exa/get_code_context_exa tools that directly control the agent's tool surface and execution, so remote code is fetched and executed and the skill depends on it.